By Sana Qureshi · Reviewed by Daniel Ruiz, CPA
This article addresses U.S. federal tax rules only. State tax treatment of digital assets varies and is not covered here.
Every crypto holder who moved coins between an exchange and a personal wallet before 2025 is sitting on a basis problem they may not know they have. The IRS ended the old “universal” cost-basis shortcut, and the replacement rule ties every unit’s basis to the specific wallet or account where it lives.
Wallet-by-wallet accounting requires tracking cost basis separately for each wallet, exchange account, or address you hold digital assets in, rather than pooling all your crypto together under one basis calculation. Under Rev. Proc. 2024-28, taxpayers who used the old universal method had to allocate their existing “unused basis” to specific wallets using a safe-harbor election completed by January 1, 2025, and every disposal after that date now defaults to first-in-first-out within each wallet unless you specifically identify units.
Not tax advice. This article is informational and general in nature. It does not replace individualized guidance from a qualified tax professional. Rules around digital asset basis allocation are technical and fact-specific. Talk to a CPA or tax attorney before making an election or filing a return that relies on it.
What Changed, and Why It Matters Now
For years, many crypto holders and even some software platforms treated all of a person’s digital asset holdings as one giant pool. If you bought Bitcoin on three exchanges and moved coins between two personal wallets, your basis tracking often ignored where each unit actually sat. That approach is what practitioners call the universal or multi-wallet method, and the IRS effectively ended it for tax years beginning after 2024.
Final regulations require basis to be determined on an account-by-account (wallet-by-wallet) basis for digital assets, aligning crypto with how brokerage accounts have long tracked stock basis. The catch: a huge number of taxpayers had already built up basis under the old pooled method, and that basis wasn’t tied to any particular wallet. Rev. Proc. 2024-28, released June 28, 2024, gave those taxpayers a one-time safe harbor to sort out the mess before the new rules took hold.
The safe harbor’s snapshot date is January 1, 2025, defined in the guidance as the moment immediately after the close of December 31, 2024. Whatever digital assets you held at that instant, in whatever wallets, became the baseline for allocating your pre-existing “unused basis.” Miss that window without a reasonable allocation, and you risk having basis assigned in ways you didn’t choose, which can inflate reported gains later.
The regulatory path here took about two years to settle. Treasury and the IRS issued proposed regulations on digital asset broker reporting on August 29, 2023, then finalized them on June 28, 2024, the same day Rev. Proc. 2024-28 came out. The expanded definition of “digital asset” for reporting purposes traces back further still, to a statutory change effective for transactions on or after January 1, 2023. None of that history changes what you owe. It does explain why so many holders only heard about wallet-by-wallet requirements recently: the rules affecting 2025 conduct weren’t finalized until the middle of 2024.
The Decision You Actually Have to Make
Strip away the jargon and there are really two decisions layered on top of each other. The first is a one-time, backward-looking choice: how do you allocate the basis you already accumulated under the old pooled method across the wallets where your coins now sit? The second is an ongoing, forward-looking choice: going forward, will you rely on default FIFO ordering inside each wallet, or will you do the work of specific identification?
Both choices are binding once made. The IRS guidance states that allocations under the safe harbor are treated as adjustments to basis under Section 1012 and are irrevocable. That’s a meaningfully different posture than most tax elections, which can sometimes be revisited on an amended return. Here, get it wrong and you generally live with it.
Timing matters differently depending on which method you pick. A global allocation (a standing rule like “assign my highest-cost lots first”) had to be established before January 1, 2025, though it can be finalized later. A specific unit allocation, where you hand-pick which exact units land in which wallet, has more breathing room: it must be completed before the earlier of your first digital asset sale on or after January 1, 2025, or your 2025 tax return due date including extensions.
The Factors That Should Actually Drive Your Choice
Three practical factors separate the people for whom this is a five-minute decision from the people who need a spreadsheet and possibly a professional.
How many wallets and exchanges are involved
Someone with a single Coinbase account and nothing else has almost nothing to allocate: there’s only one place for basis to live. Someone with four exchange accounts, two hardware wallets, and a DeFi position has real allocation decisions to make, because the same coin type sitting in different places can carry very different embedded gains depending on where the high-cost or low-cost lots end up.
How messy your transfer history is
Moving coins between your own wallets doesn’t create a taxable event, but it does create a paper trail you need to reconstruct. If you can pull clean records showing which coins moved where and when, specific identification becomes realistic. If your transfer history is a tangle of half-remembered wallet-to-wallet moves across three years, a global rule is often the more honest and defensible option.
The quality of your existing recordkeeping
This is the factor people underrate. Specific unit allocation only works if you can actually document acquisition dates, per-unit cost, and which units remain in which wallet. Crypto tax software helps, but it’s not infallible, particularly for older transactions pulled from exchanges that have since shut down or changed their export formats. Be honest about what you can actually prove, not what you assume you can prove.
My honest read, after going through this framework repeatedly: unless you’re already several months into disciplined, per-transaction logging, global allocation is the more realistic default for most holders juggling more than two or three wallets. Specific unit allocation sounds more precise, and it can be, but it only pays off if the underlying records are clean enough to survive a follow-up question from the IRS or a preparer. A defensible global rule beats an unsupportable specific-unit claim every time.
Does This Apply to NFTs and Other Token Types?
Rev. Proc. 2024-28 speaks in terms of “digital assets” broadly, which under the applicable Treasury regulations covers a wide range of blockchain-recorded value, not just Bitcoin and Ether. The safe harbor’s core conditions apply regardless of token type: the unused basis must come from units that were capital assets in your hands, and it can only be allocated to remaining holdings of the same digital asset type within a wallet, not swapped across different assets.
Where this gets genuinely murky is with less liquid holdings, such as NFTs bought and held individually or governance tokens with irregular trading histories. If you’re unsure whether a specific token category fits cleanly within the safe harbor’s terms, that’s a case for professional guidance rather than a guess, since the guidance doesn’t spell out every edge case explicitly.
How This Connects to Form 1099-DA Broker Reporting
Wallet-by-wallet accounting doesn’t exist in isolation. It’s the taxpayer-side half of a broader shift toward broker-reported digital asset transactions. Brokers, including many centralized exchanges, are being phased into reporting gross proceeds from digital asset sales on the new Form 1099-DA, with reporting obligations tied to transactions occurring in 2025 and beyond.
Basis reporting by brokers is being introduced more gradually than proceeds reporting. That gap matters practically: even once your exchange starts issuing 1099-DA forms, the basis figure on that form may only reflect activity the broker itself observed, not coins you deposited from elsewhere. Your own wallet-by-wallet records remain the backstop. A form from an exchange is a data point, not a substitute for your own documentation.
Comparing the Three Approaches
Here is how the main paths compare once you’re past the one-time transition and into ongoing wallet-by-wallet accounting.
| Approach | How it works | Best fit | Main risk |
|---|---|---|---|
| Default FIFO, no election | You take no allocation action; basis flows in first-in-first-out order within each wallet by default. | Single-wallet holders, simple histories, small balances. | Legacy pooled basis can land in the wrong wallet, overstating gains on later sales. |
| Global (wallet-by-wallet) allocation | You set one standing ordering rule before the deadline and apply it consistently across every wallet holding that asset. | Multiple wallets or exchanges, decent but imperfect records, want a defensible middle ground. | Less precise than unit-level tracking; you can’t cherry-pick lots after the fact. |
| Specific unit allocation / identification | You identify exact acquisition lots and attach them to the wallet where the corresponding units sit, transaction by transaction. | Meticulous recordkeepers, high-value holdings, active traders who want tax-efficient lot selection. | Labor-intensive; unsupportable identification can be disregarded on audit, defaulting you back to FIFO. |
A Practical Way to Picture the Trade-Off
The three approaches sit on a rough scale between effort and precision. This isn’t an IRS-published metric. It’s an editorial way to visualize the trade-off holders describe when they compare the paths.
All three paths measure their allocation against holdings as of this date; it doesn’t move regardless of which approach you pick.
A Worked Example: Choosing and Applying an Approach
Consider a hypothetical holder, call her Mara, who bought Bitcoin steadily from 2019 through 2023 on one exchange, then moved portions to a hardware wallet and a second exchange over time. Under the old universal method, her software tracked total basis across everything as one number: roughly $42,000 of basis against 1.2 BTC total, regardless of which wallet held which coins.
Come January 1, 2025, Mara has 0.5 BTC in her original exchange account, 0.4 BTC in her hardware wallet, and 0.3 BTC on the second exchange. None of that $42,000 in basis is formally attached to any one of those three locations under the old records: it’s “unused basis” in the language of Rev. Proc. 2024-28, and she has to allocate it.
Mara decides a global allocation fits her situation better than hunting down every individual lot, because her transfer history has some gaps from a wallet migration in 2021. Before the deadline, she documents a standing rule: allocate her highest-cost-basis units first to whichever wallet holds the corresponding remaining coins, in the order the coins were acquired. Applying that rule, her software assigns roughly $17,500 of basis to the exchange account, $14,000 to the hardware wallet, and $10,500 to the second exchange, proportional to both the size of each wallet’s holdings and the ordering rule she picked.
Six months later, Mara sells 0.2 BTC from her hardware wallet. Because her allocation is already locked in, that sale draws only from the $14,000 basis pool tied to that wallet, not from the exchange account’s basis, and not from a blended average across all three locations. If the sale price per BTC works out to a gain, it’s calculated using that wallet-specific basis alone. Had she done nothing and let a default FIFO rule apply retroactively without any allocation, the same sale might have drawn on a different, unintended slice of her basis history, potentially producing a different reported gain.
Compare that with a second hypothetical holder, Dev, who kept meticulous per-exchange transaction logs going back to 2017 and knows exactly which acquisition lots sit in which of his two wallets. Dev opts for specific unit allocation instead. Rather than a single ordering rule, he attaches particular lots (say, the 0.15 BTC he bought in March 2020 at a low price) directly to the wallet where those specific coins still sit, and separately identifies a higher-cost lot from 2022 sitting in his other wallet. Because his records support it down to the transaction, he can complete this allocation any time before his first 2025 sale or his extended 2025 filing deadline, whichever comes first, rather than locking in a rule before January 1, 2025 the way Mara did. The trade-off: if his records had gaps, this approach would fall apart faster than Mara’s simpler rule would.
This is also where wallet-by-wallet allocation connects to a broader compliance picture. The allocation choice Mara makes here feeds directly into how she satisfies the crypto cost basis reporting requirement on her tax return, since the basis she reports for that sale has to trace back to a documented, wallet-specific number rather than an estimate.
Common Mistakes People Make Here
A few patterns show up repeatedly among holders trying to sort this out on their own.
Treating exchange-provided cost basis as automatically correct. Exchanges only know what happened on their own platform. If you deposited coins that were acquired elsewhere, the exchange’s basis figure for those units is often wrong or missing entirely. In my assessment, this is the single most avoidable mistake in the entire process, and the one that costs people the most once an exchange’s default number gets carried straight onto a tax return unchecked.
Assuming crypto tax software has already handled the transition for you. Some platforms built tools to help with the 2024-28 allocation, but the underlying judgment call, which method to use and how to apply it, is still yours to make and document. Software can execute a rule; it can’t decide which rule fits your situation.
Waiting until the tax return is due to think about allocation. Global allocation rules needed to be set before January 1, 2025. Showing up in March 2026 without having documented anything puts you outside the safe harbor’s terms entirely.
Assuming the allocation is optional if you don’t trade often. Even infrequent sellers need a defensible basis number the moment they sell anything. Skipping the allocation doesn’t avoid the requirement. It just means reconstructing it under pressure, later, with less documentation available.
Mixing up the one-time transition allocation with ongoing accounting method choice. These are related but separate. The transition allocation deals with legacy basis as of the 2025 snapshot. Your accounting method choice governs every sale after that, wallet by wallet, going forward.
Losing transfer records for wallet-to-wallet moves. A transfer between your own wallets isn’t a sale, but if you can’t show which coins moved and when, you can’t support a specific unit allocation and may be forced back to a cruder method by default.
Action Checklist
- Pull a complete transaction history from every exchange and wallet you’ve used, including ones you no longer actively use.
- Identify how many separate wallets or accounts currently hold each digital asset type.
- Decide, honestly, whether your records support specific unit identification or point you toward a global rule.
- Document your allocation method and the date you established it. Don’t rely on memory.
- Confirm the allocation is complete before your next taxable disposal, or before your 2025 return’s extended due date, whichever applies to your method.
- Keep the underlying records (acquisition dates, per-unit cost, wallet locations) somewhere durable, not just inside software you might switch away from.
- Loop in a CPA or crypto-focused tax preparer if you hold assets across more than two or three wallets, or if any single position carries a large embedded gain.
Key Takeaways
- Wallet-by-wallet accounting ties cost basis to the specific wallet or account holding each unit, replacing the older pooled/universal approach.
- Rev. Proc. 2024-28 gave a one-time safe harbor, with a January 1, 2025 snapshot date, to allocate legacy basis into wallets.
- Global allocation rules needed to be set before that date; specific unit allocation has more time but demands better records.
- Allocations made under the safe harbor are generally irrevocable once completed.
- Doing nothing doesn’t remove the requirement. It just means a default outcome you didn’t choose.
Frequently Asked Questions
What is wallet-by-wallet accounting for crypto taxes?
It’s the requirement that cost basis be tracked separately for each wallet, exchange account, or address, rather than pooled across all of a taxpayer’s digital assets. It replaced the older universal accounting approach for tax years beginning after 2024.
Do I have to make an election under Rev. Proc. 2024-28?
Only if you previously used a universal or pooled cost-basis method and still hold digital assets acquired under that method. If you’ve always tracked basis per wallet, there may be no legacy basis to allocate.
What happens if I miss the January 1, 2025 safe-harbor date?
You can still allocate basis, but you fall outside the safe harbor’s protection. A default method, generally FIFO within each wallet, would likely apply instead, and you lose the ability to make a considered, documented choice.
Can I switch allocation methods later if I change my mind?
Generally no. The guidance treats safe-harbor allocations as binding basis adjustments under Section 1012. Reconsider carefully before finalizing, since there’s typically no do-over.
Does moving crypto between my own wallets trigger a taxable event?
No. A transfer between wallets you control isn’t a sale or exchange. It does, however, require careful recordkeeping so the basis travels with the coins to their new location.
How does this interact with Form 1099-DA reporting from brokers?
Broker reporting of digital asset sales on Form 1099-DA is being phased in for transactions starting in 2025, with basis reporting following. Your own wallet-by-wallet records remain important because brokers can only report what happened on their platform, not basis carried over from elsewhere.
Is global allocation or specific unit allocation better for most people?
Neither is universally better. Global allocation tends to suit holders with multiple wallets and imperfect records, since it requires one documented rule rather than lot-by-lot proof. Specific unit allocation suits holders with clean, detailed records who want more control over which basis attaches to which coins.
Does the safe harbor apply if I never used a universal cost-basis method?
If your records already tracked basis separately by wallet or account before 2025, there’s typically no legacy “unused basis” to allocate, and the safe harbor’s transition relief has little relevance to you. It exists specifically for taxpayers moving off the old pooled approach.
Where This Leaves You
None of this requires perfection on day one. It requires a documented, reasonable choice, made honestly and based on records you can actually support, and then consistency after that. The holders who struggle later aren’t usually the ones who picked global allocation instead of specific identification, or vice versa. They’re the ones who never made a documented choice at all and are now trying to reconstruct one from memory during filing season.
References
- Internal Revenue Service, Rev. Proc. 2024-28 (released June 28, 2024), providing transition safe harbor for digital asset basis allocation.
- The Tax Adviser, “Universal accounting for digital assets concludes, but safe harbor available,” October 2024.
- Aprio, “Overview of Rev. Proc. 2024-28: What Digital Asset Users Need to Know.”
- EisnerAmper, “Taxpayers Have Until January 1, 2025, to Make Use of Safe Harbor Digital Asset Basis Allocation,” December 2024.
